Shipping an enterprise project with Claude and Codex: AI writes, judge scores, you decide
kmeanskaran · x · 2026-11-03
The author recounts building an enterprise project on a tight deadline with Claude and Codex, arguing software engineering isn't dead—it just changed.
Key points
- Models like Opus 5.5 and GPT-6.1 Sol code so well (understanding codebases without a README) that non-technical stakeholders now expect near-zero-bug delivery in days: "You have Claude, right? Can you deliver by next Friday?"
- Solo delivery is possible but far from easy; the pressure lands on the developer.
Workflow revealed in the diagrams
- RAG pipeline: thousands of PDFs → LLM ingests tables/images → vector store with chunks+metadata → retriever with JSON lookup → LLM answers with citations; one API, 60s per answer.
- Agent loop: skills/rules → Claude/Codex agent → a judge scores against 95 Q&As, failures loop back → human reviews the diff ("AI never pushes") → GitHub Actions CI/CD → ship with Docker + Terraform.
- Deploy stack: Actions runs test/eval/build, Docker Comose hosts FastAPI app, frontend, database with Prometheus/Grafana monitoring; Terraform manages server, network, secrets.
Bottom line: AI writes, the judge scores, you decide.
More from coding & agent
- 20-year language dream handed to Opus: 200 pages of notes turned into an executable plan — pbaylies · 2026-11-03
- Paradev AI sidebar now agent-driven, surfacing diffs, files, browser and mobile app views — massimosgrelli · 2026-11-03
- DeepSeek Harness fever: two WeChat groups fill to 200 members in 10 minutes — vista8 · 2026-11-03
- Send an article link to your Claude dot and it builds a responsibilities page — craigsdennis · 2026-11-03
- Skill++ paper: routing-aware skill refinement lifts agent routing from 36.4% to 76.3% — dair_ai · 2026-11-03
- WhatsApp MCP pager calls you when your coding agent needs input — Zakariaoufi · 2026-11-03